llms.txt Content
# Zenrows
> Zenrows is web data infrastructure built for agents and applications. It gives them reliable access to the live web, including the protected web: sites defended by anti-bot systems, rendered dynamically in JavaScript, or built to block ordinary tools. Available over one API, a CLI, an MCP server, and SDKs, built around four primitives: Fetch, Extract, Batch, and Browser Sessions.
The wedge is reliable access to the protected web: Zenrows keeps working when a site is dynamic, defended by anti-bot systems, high-volume, or business-critical, where plain HTTP clients and generic scraping tools get blocked. Proxies, JavaScript rendering, and session handling are components of the platform, not the product itself: fetching a page is one primitive among several (Fetch, Extract, Batch, Browser Sessions), all reachable from a single API, CLI, MCP server, or SDK.
Agents can set up Zenrows directly from `https://www.zenrows.com/SKILL.md` (setup steps, MCP tool list, API parameters) or run `npx -y @zenrows/cli init` from a terminal. Get a free API key at `https://app.zenrows.com/register`, no credit card required.
## Platform
- [Platform overview](https://www.zenrows.com/products/): Compare Fetch, Extract, Batch, and Browser Sessions side by side, with a decision matrix and real composition examples.
- [Fetch](https://www.zenrows.com/products/fetch/): One REST call returns any page as clean HTML, Markdown, text, or PDF, handling anti-bot, JavaScript rendering, geo-targeting, and sessions automatically.
- [Extract](https://www.zenrows.com/products/extract/): Turn any page into structured JSON in a single request, via autoparse, CSS selectors, or clean Markdown.
- [Batch](https://www.zenrows.com/products/batch/): Push thousands of URLs as one job; workers drain in parallel and deliver JSON per URL via webhook.
- [Browser Sessions](https://www.zenrows.com/products/browser-sessions/): A real, persistent browser session for logins, clicks, and multi-step flows that